For monolingual individuals eager to reach out to global audiences, the barriers to communication have dramatically lowered. Reliable services like Google Translate offer the ability to translate images, audio, and complete websites across numerous languages. Meanwhile, newer applications such as ChatGPT are emerging as convenient mobile translation aids.
On the technological forefront, companies like DeepL and ElevenLabs have achieved impressive billion-dollar valuations through advanced language technologies that businesses can integrate into their applications. Enter a new contender in this space: an AI-driven localization platform designed to empower developers to go global—essentially, a “Stripe” for app localization.
Previously known as Replexica, Lingo.dev focuses on developers seeking to make their applications fully localized from the start. Developers only need to continue coding as usual, with Lingo.dev seamlessly handling translation processes in the background. This alleviates the hassle of copying and pasting text into tools like ChatGPT for quick translations or dealing with numerous translation files in various formats from different providers.
Currently, Lingo.dev serves clients including the French unicorn Mistral AI and Cal.com, an open-source alternative to Calendly. To facilitate its next growth phase, the startup recently announced a successful $4.2 million seed round led by Initialized Capital, with additional backing from Y Combinator and various angel investors.
Found in Translation
Lingo.dev is the brainchild of CEO Max Prilutskiy and CPO Veronica Prilutskaya (as shown above). The duo previously sold their SaaS startup, Notionlytics, to an undisclosed buyer last year. They had been laying the groundwork for Lingo.dev since 2023, with their initial prototype developed during a Cornell University hackathon. This led them to their first clients and a spot in Y Combinator’s fall program last year.
At its core, Lingo.dev operates as a Translation API that can be accessed locally by developers either via their CLI (command line interface) or through direct integration with CI/CD systems such as GitHub or GitLab. In essence, development teams receive automated translation updates through pull requests whenever standard code changes occur.
Central to this process is a large language model (LLM), or more accurately, several LLMs, which Lingo.dev orchestrates to manage various inputs and outputs. This hybrid strategy, leveraging models from providers such as Anthropic and OpenAI, ensures that the most suitable model is selected for a given task.
“Different prompts yield better results with certain models than with others,” Prilutskiy shared with TechCrunch. “Also, depending on the specific use case, we may prioritize lower latency—or latency may not be a concern at all.”
Discussing LLMs inevitably brings up concerns about data privacy—one reason some businesses are hesitant to adopt generative AI. Lingo.dev prioritizes the localization of front-end interfaces, although it also accommodates business content like marketing websites and automated emails. Importantly, it does not process any clients’ personally identifiable information (PII).
“We do not anticipate any personal data being transmitted to us,” stated Prilutskiy.
Through Lingo.dev, businesses can create translation memories (archives of previous translations) and upload their style guides to ensure brand consistency across different markets.
Companies can also define how certain phrases should be treated and in which contexts. The localization engine can assess the placement of specific text, making necessary modifications as needed—for example, when translating from English to German, a word may double in length and disrupt the user interface (UI). Users can direct the system to rephrase sentences to maintain the original length.
Without contextual insight into the application itself, localizing isolated pieces of text—like a label—can pose challenges. Lingo.dev addresses this through a feature called “context awareness,” which evaluates the total content of the localization file, including surrounding text or system keys that translation files may contain. It’s about grasping the “microcontext,” as Prilutskiy describes it.
The future holds even more advancements.
“We are currently developing a new feature that utilizes app UI screenshots to extract additional contextual information about UI elements and their intended use,” Prilutskiy stated.
Going Local
Lingo.dev is still in the early stages of its journey towards comprehensive localization. For instance, different colors and symbols may convey varying meanings in diverse cultures, an aspect not yet directly incorporated into Lingo.dev’s offerings. Additionally, variations in measurement systems, such as metric versus imperial, must still be handled at the developer level.
However, Lingo.dev does integrate with the MessageFormat framework, which addresses differences in pluralization and gender-specific wording across languages. Recently, the company also introduced an experimental beta feature aimed at translating idioms. For example, “to kill two birds with one stone” has a German counterpart that essentially means “to hit two flies with one swat.”
In addition to these features, Lingo.dev is conducting applied AI research to enhance various aspects of automated localization.
“One of the intricate challenges we’re currently tackling is how to preserve masculine and feminine versions of nouns and verbs during translations,” Prilutskiy explained. “Languages convey varying levels of information. For instance, the English word ‘teacher’ is gender-neutral, whereas in Spanish, it’s either “maestro” (male) or “maestra” (female). Ensuring these subtleties are accurately reflected is a significant focus of our applied AI research.”
Ultimately, the vision transcends mere translation; the aim is to approximate the level of quality one might expect from a team of professional translators.
“Overall, the [goal] with Lingo.dev is to eliminate friction in the localization process to such an extent that it seamlessly integrates as an infrastructure layer within the tech stack,” Prilutskiy concluded. “This is reminiscent of how Stripe redefined online payments, effectively becoming a fundamental toolkit for developers.”
Although the founders were initially based in Barcelona, they are relocating their company headquarters to San Francisco. Lingo.dev currently has just three employees, including the founding engineer, embodying a lean startup approach they intend to maintain.
“Everyone at YC, along with myself and other founders, strongly believe in this method,” noted Prilutskiy.
Their previous venture, which offered analytics for Notion, was entirely bootstrapped and served notable clients such as Square, Shopify, and Sequoia Capital—all without a single employee aside from Max and Veronica.
“We were just the two of us working full-time, with the occasional contractor for specific tasks,” Prilutskiy remarked. “We know how to accomplish a lot with minimal resources. Our past experience shaped this, and we are replicating the same lean strategy—now with additional funding.”
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
